[HYPERCUBE] add real region names. improve sampling of massif names for the fast mode

experiment/meteo_france_data/scm_models_data/abstract_extended_study.py

@@ -9,11 +9,13 @@ from utils import classproperty
 
 class AbstractExtendedStudy(AbstractStudy):
 
-
-
     @classproperty
     def region_names(cls):
-        return ['Alps', 'Northern Alps', 'Central Alps', 'Southern Alps', 'Extreme South Alps']
+        return ['Alps'] + cls.real_region_names
+
+    @classproperty
+    def real_region_names(self):
+        return ['Northern Alps', 'Central Alps', 'Southern Alps', 'Extreme South Alps']
 
     @property
     def nb_region_names(self):

experiment/meteo_france_data/scm_models_data/visualization/study_visualization/study_visualizer.py

@@ -10,6 +10,7 @@ import numpy as np
 import pandas as pd
 import seaborn as sns
 
+from experiment.meteo_france_data.scm_models_data.abstract_extended_study import AbstractExtendedStudy
 from experiment.trend_analysis.abstract_score import MeanScore, AbstractTrendScore
 from experiment.meteo_france_data.scm_models_data.abstract_study import AbstractStudy
 from experiment.trend_analysis.univariate_test.abstract_gev_change_point_test import AbstractGevChangePointTest
@@ -414,9 +415,17 @@ class StudyVisualizer(VisualizationParameters):
     @cached_property
     def _massif_name_to_gev_change_point_test_results(self):
         massif_name_to_gev_change_point_test_results = {}
-        massif_names = self.study.study_massif_names
-        if self.nb_massif_for_change_point_test is not None:
-            massif_names = sample(massif_names, self.nb_massif_for_change_point_test)
+        if self.nb_massif_for_change_point_test is None:
+            massif_names = self.study.study_massif_names
+        else:
+            # Get one massif from each region to ensure that the fast plot will not crash
+            assert self.nb_massif_for_change_point_test >= 4, 'we need at least one massif from each region'
+            massif_names = [AbstractExtendedStudy.region_name_to_massif_names[r][0]
+                            for r in AbstractExtendedStudy.real_region_names]
+            massif_names_for_sampling = list(set(self.study.study_massif_names) - set(massif_names))
+            nb_massif_for_sampling = self.nb_massif_for_change_point_test - len(AbstractExtendedStudy.real_region_names)
+            massif_names += sample(massif_names_for_sampling, k=nb_massif_for_sampling)
+
         for massif_id, massif_name in enumerate(massif_names):
             years, smooth_maxima = self.smooth_maxima_x_y(massif_id)
             gev_change_point_test_results = compute_gev_change_point_test_results(self.multiprocessing, smooth_maxima,